14263
Blog
Andy Pollen

The Foundation Model Lexicon

aiOS, Care Coordination, Foundation Models, Workflow
Table of Contents

Foundation models are reshaping clinical AI, expanding it from narrow, single-use tools into end-to-end, multimodal intelligence. For health systems, that shift is more than technical — it changes how physicians practice medicine and how leaders evaluate strategies, guide teams and plan for the future.

This post breaks down the essential terms and concepts so you can cut through the jargon, understand what’s at stake and speak with confidence about the technology defining the next era of clinical AI.

Agentic AI

AI systems that proactively initiate context-aware actions such as analyzing patient cases for all relevant diseases, measuring and characterizing findings, prioritizing suspected urgent cases, triggering workflows or coordinating care teams without step-by-step human instruction.
Why it matters: Moves AI from passive to active decision support, accelerating care and reducing delays that can cost patients and providers.

Aidoc aiOS™

Aidoc’s enterprise-grade clinical AI operating system is the platform that orchestrates multiple AI models — Aidoc’s, third-party and homegrown — integrating them seamlessly into workflows with governance, monitoring and scalability.
Why it matters: aiOS is the only scalable AI platform that can integrate into native systems. It organizes and delivers Aidoc’s CARE™ foundation model insights into clinical workflows. 

Clinical AI Reasoning Engine (CARE)

Aidoc’s clinical-grade foundation model, trained on real-world, multimodal data. It already powers FDA-cleared applications across multiple domains, enabling faster, more generalizable AI solutions with consistent and unmatched performance.
Why it matters: From a single foundation, health systems can fine-tune Aidoc’s CARE™ for tasks ranging from triage and detection to measurement, reporting and prediction.

Contrastive Language-Image Pre-Training (CLIP)

A training technique that’s critical to self-supervised learning, it teaches AI to connect images with descriptive text, enabling shared understanding between visual and language data.
Why it matters: Strengthens multimodal reasoning, allowing AI to “see” and “describe” findings in clinically meaningful ways.

Deep Learning

A method where the model learns to make decisions by training on large sets of labeled data, using layered neural networks to recognize complex patterns — like spotting disease in medical images.
Why it matters: The backbone of early clinical AI (supervised learning), it’s now limited in scope compared to foundation models.

Domain

A clinical area or problem space the model is trained on — such as radiology, cardiology or lab data.
Why it matters: Domains define where AI can deliver measurable outcomes. Foundation models allow expansion across domains without rebuilding from scratch.

Fine-Tuning

Adapting a pre-trained model to a specific task using a small, labeled dataset, enabling fast and efficient training.
Why it matters: Delivers high accuracy AI at a lower cost and faster development timelines.

Foundation Model

A large, pre-trained model that learns general data representations and can be adapted to many clinical tasks.
Why it matters: Provides the base layer of intelligence for scalable AI, enabling broad coverage and stronger performance than task-specific models.

Generalizable

The ability of an AI model to perform well across many tasks and datasets — not just those it was originally trained on.
Why it matters: Turns AI from a narrow tool into a system-wide capability, giving leaders confidence it will perform across hospitals, populations and workflows.

Multimodality

The ability to learn from and combine different types of data — like images, text and structured electronic health records (EHRs) — to improve clinical insight.
Why it matters: Preventable care gaps often stem from siloed data; multimodal AI integrates signals for a more complete patient view.

Neural Network

A layered structure of algorithms inspired by the brain, used to recognize patterns in data.
Why it matters: The fundamental building block of modern AI but only as powerful as the data and infrastructure surrounding it.

Self-Supervised Learning

A method where the model learns from unlabeled data by solving pretext tasks with known answers, like predicting missing parts of an image.
Why it matters: Unlocks the ability to train on vast, real-world data without the bottleneck of manual labeling.

Supervised Learning (Traditional AI Training)

An approach where the model learns from manually labeled data, one task at a time.
Why it matters: The foundation of early AI development but too slow and narrow to meet enterprise demands today.

Want to go deeper?

Explore our Foundation Model Resource Hub for infographics, videos and interviews that show how foundation models are shaping the future of clinical AI — and what it means for your health system today.

Explore the Latest AI Insights, Trends and Research

Follow a manual added link

Webinars

Aidoc experts, customers and industry leaders share the latest in AI benefits and adoption.

Follow a manual added link

eBooks

Explore how clinical AI can transform your health system with insights rooted in real-world experiences.

Follow a manual added link

AI Strategy

Learn how to go beyond the algorithm to develop a scalable AI strategy and implementation plan.

Follow a manual added link

Solutions

Explore how Aidoc can help increase hospital efficiency, improve outcomes and demonstrate ROI.

Andy Pollen
Andy Pollen is an experienced healthcare communicator and strategist who currently serves as the Director of Marketing Communications for Aidoc. Previously, he was the global marketing communications lead for critical care solutions within 3M Health Care's Medical Solutions Division, now Solventum. Pollen has also held communications positions with the University of Minnesota Academic Health Center, Indiana University Health and several business functions within Eli Lilly and Company through Borshoff, a creative services agency. He earned a bachelor’s degree in public relations and journalism from Ball State University and holds a master’s degree in business administration from Anderson University.
Andy Pollen
Director, Marketing Communications

© Copyright 2025 Aidoc